I refer to this question's answer, the second part mentioning the random data, and no headers or anything, including files that are recognizable. Reading the answer just made me think adding random file data would just further the security?
If an encryption method does make all of the data (when viewing the HEX values) random, wouldn't it add an extra level of security to have the encryption algorithm add the values of random, well-known file types, in random locations, and selecting the type from predetermined well-known file types (.PNG, . Jpeg,.TXT and more) and placing a random amount of random file types (randomly selected from the list or known file types/extensions), in random places. Maybe making it not place them in areas that usually contain headers and other information. So it would be not 100% random, because you need to know the types of files that are added, and a possible pseudo-random algorithm in how frequent or how many are placed, simply because you cannot have the randomness producing too many of these fake extensions, and don't want them all over the place, I assume that wouldn't be good? Just based on memory?
I feel like if you can add a certain degree of questioning from whomever is examining it, that it isn't all random, that would add security because they could not determine it is random data, because the traces of files. Ultimatley making it just harder to determine what is on a hard drive or SDD or even files.